Download Simultaneous Multiparametric PET/MRI with Silicon Photomultiplier

Downloaded from jnm.snmjournals.org by Dong Soo Lee on November 8, 2016. For personal use only.
Simultaneous Multiparametric PET/MRI with Silicon
Photomultiplier PET and Ultra-High-Field MRI
for Small-Animal Imaging
Guen Bae Ko1,2, Hyun Suk Yoon1,2, Kyeong Yun Kim1,2, Min Sun Lee1,3, Bo Yeun Yang1,4, Jae Min Jeong1–4,
Dong Soo Lee1,3–5, In Chan Song4,6, Seok-ki Kim7,8, Daehong Kim8, and Jae Sung Lee1–4
1Department
of Nuclear Medicine, Seoul National University, Seoul, Korea; 2Department of Biomedical Sciences, Seoul National
University, Seoul, Korea; 3Interdisciplinary Program in Radiation Applied Life Science, Seoul National University, Seoul, Korea;
4Institute of Radiation Medicine, Medical Research Center, Seoul National University, Seoul, Korea; 5Department of Molecular
Medicine and Biopharmaceutical Sciences, Graduate School of Convergence Science and Technology, Seoul National University,
Suwon, Korea; 6Department of Radiology, Seoul National University, Seoul, Korea; 7Department of Nuclear Medicine, National
Cancer Center, Goyang, Korea; and 8Molecular Imaging and Therapy Branch, National Cancer Center, Goyang, Korea
Visualization of biologic processes at molecular and cellular levels
has revolutionized the understanding and treatment of human
diseases. However, no single biomedical imaging modality provides
complete information, resulting in the emergence of multimodal
approaches. Combining state-of-the-art PET and MRI technologies
without loss of system performance and overall image quality can
provide opportunities for new scientific and clinical innovations. Here,
we present a multiparametric PET/MR imager based on a smallanimal dedicated, high-performance, silicon photomultiplier (SiPM)
PET system and a 7-T MR scanner. Methods: A SiPM-based PET
insert that has the peak sensitivity of 3.4% and center volumetric
resolution of 1.92/0.53 mm3 (filtered backprojection/ordered-subset
expectation maximization) was developed. The SiPM PET insert was
placed between the mouse body transceiver coil and gradient coil of
a 7-T small-animal MRI scanner for simultaneous PET/MRI. Mutual
interference between the MRI and SiPM PET systems was evaluated
using various MR pulse sequences. A cylindric corn oil phantom was
scanned to assess the effects of the SiPM PET on the MR image
acquisition. To assess the influence of MRI on the PET imaging functions, several PET performance indicators including scintillation pulse
shape, flood image quality, energy spectrum, counting rate, and
phantom image quality were evaluated with and without the application of MR pulse sequences. Simultaneous mouse PET/MRI studies
were also performed to demonstrate the potential and usefulness of
the multiparametric PET/MRI in preclinical applications. Results: Excellent performance and stability of the PET system were demonstrated, and the PET/MRI combination did not result in significant
image quality degradation of either modality. Finally, simultaneous
PET/MRI studies in mice demonstrated the feasibility of the developed system for evaluating the biochemical and cellular changes in
a brain tumor model and facilitating the development of new multimodal imaging probes. Conclusion: We developed a multiparametric
imager with high physical performance and good system stability and
Received Nov. 19, 2015; revision accepted Mar. 11, 2016.
For correspondence or reprints contact either of the following:
Jae Sung Lee, Department of Nuclear Medicine, Seoul National University,
103 Daehak-ro, Jongno-gu, Seoul 110-799, Korea.
E-mail: [email protected]
Daehong Kim, Molecular Imaging and Therapy Branch, National Cancer
Center, Goyang 410-769, Korea.
E-mail: [email protected]
Published online Apr. 14, 2016.
COPYRIGHT © 2016 by the Society of Nuclear Medicine and Molecular
Imaging, Inc.
demonstrated its feasibility for small-animal experiments, suggesting
its usefulness for investigating in vivo molecular interactions of metabolites, and cross-validation studies of both PET and MRI.
Key Words: PET/MRI; silicon photomultiplier (SiPM); hybrid imaging;
multi-parametric imaging; dual-modality imaging probe
J Nucl Med 2016; 57:1309–1315
DOI: 10.2967/jnumed.115.170019
A
substantial role of small-animal imaging has been pinpointed
in numerous studies in terms of understanding the underlying mechanism of human diseases and elucidating the efficacy of new therapeutic approaches. Among the in vivo small-animal imaging modalities,
which are scaled down to dedicated devices from clinical ones,
PET is the most-sensitive technique that is readily translatable to
the clinic (1). Spatial and temporal distributions of compounds labeled
with a positron-emitting radionuclide are noninvasively measured by
the PET scanner. Consequently, the PET scanner provides quantitative
information about various physiologic and biochemical processes,
such as glucose metabolism, gene expression, and drug occupancy
(2–5). Additionally, insufficient morphologic information in PET
images can be compensated for by combining the PET scanner
with morphologic imaging devices, such as CT and MRI.
Despite the great success of PET/CT in both clinical and preclinical applications, a system combining PET and MRI has been
demanded because of the advantages of MRI over CT (6,7). Particularly in small-animal imaging studies, the use of MRI reduces
radiation exposure to the animal, which needs to be minimized in
longitudinal experiments to avoid confounding biologic effects (8).
A simultaneously operating PET/MRI system is advantageous in
reducing acquisition time and anesthesia dose and in attaining nearperfect spatial and temporal correlation between the information
provided by the 2 imaging modalities. In addition, the functional
capability of MRI has been dramatically enhanced in the last 2 decades,
opening new opportunities for studying pathology and biochemical
processes through the combinations of multiparametric MRI and PET.
Early attempts at simultaneous PET/MR image acquisition using
long optical fiber bundles did not provide sufficiently good PET
SIMULTANEOUS MULTIPARAMETRIC PET/MRI • Ko et al.
1309
Downloaded from jnm.snmjournals.org by Dong Soo Lee on November 8, 2016. For personal use only.
performance because of the large light loss in the fiber bundles that
transfer scintillation light photons generated from the scintillation
crystals installed inside the MRI magnet to the photomultiplier
tubes located outside the magnet (9,10). To overcome this problem,
the magnetically insensitive and compact avalanche photodiode had
been suggested as an alternative photosensor for PET detectors and
led to the successful development of preclinical and clinical simultaneous PET/MRI systems (11–14). However, the relatively low
signal amplification gain and slow time response of avalanche photodiodes compared with photomultiplier tubes have stressed the
demand for more advanced semiconductor photosensors in nextgeneration simultaneous PET/MRI systems. In recent years, several
groups, including us, have demonstrated the feasibility of smallanimal PET and PET/MR imaging based on silicon photomultipliers
(SiPMs); SiPMs are composed of an array of micro–avalanche
photodiode cells operating in Geiger mode, which provides a
much higher signal amplification gain and faster timing properties
than the avalanche photodiode (15–20).
Here, we present an SiPM-based MR-compatible PET imaging
system that is combined with a 7-T small-animal dedicated MR scanner.
The results of mutual interference evaluation and simultaneously
acquired PET/MR images from 3 different mice studies are presented.
Collectively, these studies demonstrate that the developed PET/MRI
system enables in vivo multiparametric PET/MRI studies with high PET
performance and negligible degradation in 7-T MR performance.
MATERIALS AND METHODS
MR-Compatible SiPM PET System
The SiPM is a semiconductor photosensor that offers extremely high
counting rates and photon number resolution with MR compatibility.
We designed the SiPM PET system to be placed between the mouse
volume radiofrequency coil and the actively shielded gradient coil of a
7-T small-animal MR scanner (BioSpec 70/20 USR; Bruker Biospin)
(Fig. 1; Supplemental Fig. 1 [supplemental materials are available at
http://jnm.snmjournals.org]). The radiofrequency and gradient coils have
inner/outer diameters of 35/60 and 114/198 mm, respectively. Thus, in
designing the SiPM PET insert for the 7-T MR scanner, we focused
on minimizing the thickness of the PET system to fit within the narrow
space between the coils. Nonmagnetic passive electronic components
and connectors were used in the parts placed inside the MR scanning
room. The SiPM PET consists of 16 detector modules, and each detector
module includes 4 detector blocks. Each detector block is composed of a
9 · 9 array of lutetium yttrium orthosilicate (LYSO) crystals (1.2 · 1.2 ·
10 mm3; SIPAT) and a large-area monolithic 4 · 4 SiPM array (S118283344M; Hamamatsu Photonics). Optical grease (BC-630; Oken) was
used to optically couple the LYSO crystals and SiPM arrays without
inserting any light guides between them. The 16 AC-coupled anode
signals of the SiPM array were multiplexed to 4 position signals through
a resistive charge division network (21).
To minimize heat generation from the electronics, low-power amplifiers were placed about 15 cm from the SiPM detectors. We also
implemented a real-time gain compensation module based on the
temperature monitoring and bias voltage feedback to maintain uniform performance (22). The real-time adjustment of bias voltage
yielded stable photopeak positions (,0.5% variation), energy resolutions (average energy resolution was maintained at 14.2% despite the
temperature variation), and prompt counting rates (,1.2% variation)
over a wide temperature range (10C–30C).
The PET insert was shielded by a 1-mm-thick carbon fiber tube
(inner/outer diameters, 98/100 mm), a 3-dimensional (3D) printed
front cap coated with a 30-mm copper film, and a rear cap made of
1-mm-thick aluminum for electromagnetic shielding. The thicknesses
of the shield materials, which were closely located in the PET field of
view (FOV), were carefully determined regarding their conductivity (carbon fiber, 5.88 · 104; copper, 5.95 · 107; and aluminum, 3.55 · 107 S/m)
to minimize eddy currents and perform sufficient radiofrequency blocking. The axial FOV and inner diameter of the PET insert were 55 and
60 mm, respectively (Table 1). The centers of the MR magnet and PET
axial FOV were aligned for simultaneous PET/MR imaging studies. The
temperature, applied bias voltage, and current were continuously monitored
during PET/MRI data acquisition.
The PET insert yields superb performance for small-animal imaging
(22). The radial/tangential/axial resolutions estimated from reconstructed point source images using ordered-subset expectation maximization were 0.75/0.68/1.05 mm at the center of the FOV and 1.46/0.71/
1.22 mm at the 14-mm radial off-center. When the filtered backprojection
algorithm was used, the radial/tangential/axial resolution ranged from 1.31/
1.14/1.29 mm at the center of the FOV to 2.18/2.11/1.58 mm at the 14-mm
radial off-center. The absolute peak sensitivity was 3.36%, with a coincidence time window of 12 ns and an energy window of 250–750 keV. The
peak noise equivalent counting rate and scatter fraction were 42.4 kcps
at 15.08 MBq and 16.5% with the same window setting, respectively, for
a National Electrical Manufacturers Association mouse-sized phantom.
TABLE 1
Major Characteristics of SiPM PET Insert
Characteristic
Detector face to face (mm)
1310
THE JOURNAL
OF
64
Scintillator material
LYSO
Crystal size (mm3)
1.2 · 1.2 · 10.0
Crystal pitch (mm)
1.28
No. of crystal rings
36
No. of crystals/ring
144
Total no. of crystals
5,184
Axial FOV (mm)
FIGURE 1. (A) Three-dimensional drawing of PET/MR system. (B) Picture
of PET insert inside carbon fiber shield shows detector modules, amplifiers,
biasing circuits (circ), and communication circuit.
Value
Insert length (cm)
Insert inner diameter/outer diameter (mm)
NUCLEAR MEDICINE • Vol. 57 • No. 8 • August 2016
55
82
60/99
Downloaded from jnm.snmjournals.org by Dong Soo Lee on November 8, 2016. For personal use only.
All PET and PET/MR studies were performed at 20C 6 2C with
the real-time bias voltage adjustment. PET data were acquired using a
custom-made field-programmable gate array (FPGA)–based data acquisition (DAQ) system. The differential position signals (scintillation
pulse after passing resistive charge division network) from the amplifiers were digitized by free-running analog-to-digital converters with
12-bit resolution and 125 Msps sampling speed (ADS6425; Texas
Instruments). Once a trigger signal was detected, an FPGA (Virtex6;
Xilinx) generated a time stamp with a bin size of 2 ns and stored the
streams of position signals. The baseline drift was compensated for by
subtracting DC level calculated from eight samples preceding the
rising edge of the position signal, and then the integral value of each
position signal was calculated by integration of signals for 256 ns (32
samples for 125 Msps). Single-event information containing DAQ
channel number, time stamp, and position signal integral values
were transferred to the personal computer via gigabit Ethernet. The
list-mode data were saved with the exact detector block position
decoded from the DAQ channel number and the position signal integral
values in the personal computer. Three-dimensional line-of-response
(LOR) histograms of prompt coincidence events were generated from
the list-mode data. The energy and timing windows were set to 350–
650 keV and 12 ns, respectively, unless otherwise stated. The LOR
histograms were corrected for random coincidence events estimated
from single rates. Component-based normalization was performed to compensate for sensitivity variation in LORs, and normalization components were derived by scanning cylinder and annulus phantoms filled
with 18F-FDG covering the entire FOV. All data were reconstructed
using LOR-based, 3D ordered-subset expectation maximization with 3
iterations and 12 subsets and postsmoothed using a 3D Gaussian kernel
of 1 mm in full width at half maximum.
Assessment of Mutual Interference Between PET and MRI
Influence of PET on MRI. To assess the influence of the PET insert
on MR measurement, a uniform corn oil (which has good chemical
and thermal stability and appropriate T1, T2, and proton density
values that are within the biological range) phantom with T1 and T2
relaxation times of 423 and 130 ms, respectively, was scanned. Main
magnetic field (B0) stability, image quality, and temporal stability of
image intensity were evaluated following similar procedures used in
Wehrl et al. and Hong et al. (23,24). Great care was taken to ensure
that the relative positions of the phantom, radiofrequency coil, and
main magnet were not changed during the comparative assessment.
Prior to obtaining MR scans, the proton resonance spectrum was
obtained to verify the effect of the PET insert on the resonant frequency. This procedure was necessary because the off-resonance operation manifests as a reduction in image signal-to-noise ratio (SNR),
system sensitivity, and image linearity. SNR and integral uniformity
for 4 frequently used MRI sequences in routine animal scans of
2-dimensional (2D) fast spin-echo (FSE), 2D spin-echo (SE), 2D gradientecho (GRE), and 3D spoiled gradient-recalled-echo (SPGR) (Supplemental Table 1) were analyzed using the following equations:
and the factor 0.655 in the SNR equation was introduced due to Rician
distribution of background noise in magnitude images.
B0 inhomogeneity can be caused by PET-induced magnetic susceptibility variations and regional eddy current generations, even when
appropriate magnet shimming has been used. Therefore, the B0 distortion
was also measured by a phase measurement method. A multiple gradientecho sequence with the following parameters was applied to obtain
phantom images in the coronal plane: repetition time (TR), 500 ms; first
echo time (TE1), 10 ms; second echo time (TE2), 20 ms; third echo time
(TE3), 30 ms; FOV, 35 · 60 mm2; matrix size, 64 · 128; slice thickness,
2.0 mm; number of slices, 3; and pixel bandwidth, 257 Hz/pixel. From the
phase map derived from the images obtained with TE1 [F(r, TE1)] and
TE2 [F(r, TE2)], B0 maps were generated as follows:
DB0 ðrÞ 5
Fðr; TE2 Þ 2 Fðr; TE1 Þ
;
gðTE1 2 TE2 Þ
where g is the gyromagnetic ratio of 1H.
Gradient-recalled echo planar imaging (EPI) used in functional
MRI studies was evaluated with the following parameters: TR/TE,
2,000 ms/9 ms; flip angle, 30˚; FOV, 60 · 60 mm2; matrix size, 64 · 80;
slice thickness, 1.5 mm; number of slices, 5; and pixel bandwidth, 67.9
Hz/pixel. A total of 180 volume scans were acquired over 6 min. The
mean signal intensity and percentage temporal drift of the signal (i.e.,
the standard deviation of signal intensity) on the ROIs were evaluated as
a performance indicator in the functional MRI experiments.
Influence of MRI on PET. To assess the influence of MRI on the PET
insert installed inside the 7-T magnet, several PET factors including
scintillation pulse shape, flood image quality, energy spectrum, count
rate, and phantom image quality measured with and without application
of radiofrequency sequences were compared. The 2D FSE, 3D SPGR,
and EPI sequences with identical parameters used in the assessment of
the effects of PET on MRI were used for these experiments, except for
the number of excitations (NEX) to make the scan time longer than
5 min for each acquisition. The scintillation pulses were captured with a
125-Msps analog-to-digital converter used in the DAQ system and examined for baseline distortion and radiofrequency and gradient-induced
noise. The flood map quality and energy spectrum, including photopeak
fraction and energy resolution, were also evaluated on data obtained
with an 18F-FDG filled uniform phantom. Blank PET scan data were
also acquired to investigate count rate variation based on gradient and
radiofrequency pulse applications. The single-event rates triggered by
the intrinsic activity of 176Lu of the LYSO crystal were measured in
each condition, with a threshold level of approximately 50 keV. The
uniform and hot-rod phantom filled with 18F-FDG was scanned for
30 min to investigate the artifacts in the reconstructed PET images. Care
was taken to ensure that the activity used in the simultaneous PET/MR
and standalone PET scans was not different (7.4 MBq 6 5%). The
standalone PET scan was obtained outside the MR magnet after the
radiofrequency coil was removed.
Animal Experiments
m
SNR 5 0:655 S
sN
All animal experiments were approved by the Institutional Animal
Care and Use Committee at the National Cancer Center. PET and MR
images were fused after spatial registration with a rigid-body transformation between them using AMIDE and FIRE software (25,26).
The detailed MR protocols used for animal studies are described in
Supplemental Table 2.
ðSmax 2 Smin Þ
· 100%;
Integral uniformity 5 1 2
ðSmax 1 Smin Þ
Glucose Metabolism Imaging
where mS is the mean signal intensity in a region of interest (ROI)
placed in a corn-oil-filled region (ROIs), sN is the standard deviation
of signal intensity derived from an ROI placed in a background region
(ROIN), Smax and Smin are the maximum and minimum values in ROIs,
To demonstrate the feasibility of the developed PET/MRI system in a
mouse neuroimaging study, we examined the metabolic function of the
mouse brain. We intravenously injected 18F-FDG of 17.2 MBq/0.2 mL via
the tail into a BALB/c mouse (female; age, 21 wk; weight, 19 g) under
SIMULTANEOUS MULTIPARAMETRIC PET/MRI • Ko et al.
1311
Downloaded from jnm.snmjournals.org by Dong Soo Lee on November 8, 2016. For personal use only.
light isoflurane anesthesia. After a 10-min uptake, PET was performed
for 60 min simultaneously with high-resolution T2-weighted MRI
(TR/TE, 2,500/35 ms; echo train length [ETL], 8; NEX, 16).
Tumor Imaging
U87MG glioblastoma cells were implanted into the brain of an 8-wk-old
female BALB/c nude mouse. Simultaneous imaging study was performed
on the mouse at 10 wk of age (18.4 g), 20 min after intravenous injection of 12.2 MBq of 11C-MET. MR sequences and their parameters
used in the study are as follows: FSE (TR/TE, 2,500/35 ms; ETL, 8; NEX,
4), fluid-attenuated inversion recovery (FLAIR; TR/TE, 3,000/10 ms;
inversion time, 825 ms; ETL, 2; NEX, 1), and diffusion-weighted imaging
(TR/TE, 2,000/27 ms; b values, 0, 45, 350, 1,000, and 2,000 s/mm2).
The apparent diffusion coefficient (ADC) map was calculated from the
diffusion-weighted images.
Lymph Node (LN) Imaging
Simultaneous PET/MRI is also a useful tool for developing multimodal imaging probes. An imaging study with 64Cu-NOTA-ironoxidemannose (64Cu-NOTA-IO-MAN), a new dual-modality probe yielding
superb T2 contrast in MRI (supplemental data; Supplemental Fig. 2),
was performed to demonstrate this (27). We chose the lymphatic system
as an in vivo model because of its clinical significance, particularly in
oncologic studies in which the invasion status of the LN is crucial information for disease stratification, staging, and management (28).
First, T2-weighted MR images (TR/TE, 2,500/30 ms; ETL, 4; NEX,
4) of the lower abdominal area and legs of an anesthetized BALB/c
mouse (female; age, 6 wk; weight, 18.5 g) including both left and
right popliteal LNs in the FOV were acquired. 64Cu-NOTA-IO-MAN
(0.37 MBq/10 mL) was then intradermally injected into the left hind
paw of the mouse, carefully maintaining the posture. Simultaneous
PET/MRI using the same MRI parameters as the preinjection scan was
performed 10 min and 2 h after injection. After the image acquisition,
the animal was sacrificed using CO2 inhalation, and both left and right
popliteal LNs were isolated. The isolated tissue was embedded in paraffin and dissected at 8 mm. The sections were deparaffinized and then
stained with Prussian blue to detect iron oxide particles and counterstained with nuclear fast red. The specimens were dehydrated and
mounted on a coverslip, and microscope analysis was performed.
RESULTS
Influence of PET on MR Performance
The shift of proton resonance frequency caused by the insertion
of the PET system into the MR system was less than 300 Hz (i.e.,
1.0 ppm at 7 T) regardless of the PET power on/off condition.
There were no perceptible distortions in the MR images acquired
with and without the PET insert (Supplemental Fig. 3). Additionally, there were almost no differences in the MR image intensity
levels and uniformities with and without PET for various MR
imaging sequences used in routine animal studies (Fig. 2A). However, the PET insert induced considerable SNR degradations in
MR images, mainly because of the increased variations in background areas, and the SNRs became worse (10% more degradation) when the power of the PET insert was turned on (Fig. 2A).
This result might imply that the quality factor of the radiofrequency coil was degraded by the loading effect of the PET insert
and that additional noise was introduced to the radiofrequency coil
by the penetrating electromagnetic induction generated from the
PET electronics. B0 was slightly distorted (,0.28 ppm) with the
PET insert, relative to that without the PET insert (,0.25 ppm),
because of the magnetic susceptibility of the PET component, current supplied to the detector modules, and eddy current in the shield
materials (Fig. 2B). Nevertheless, both results were well below the
maximum field variation allowed in acceptance testing of the Bruker
system (1 ppm). In addition, EPI was well preserved without quality
degradation and image distortion after the PET system was inserted
and its power turned on (Fig. 2C). The relative SD of EPI signal
intensity was 0.15% in all conditions, which was measured over 180
image sets and obtained for 6 min (Fig. 2C). This result is a solid
indication that our PET/MRI system is feasible for multifunctional
simultaneous PET/MRI studies because the typical blood oxygen
level–dependent signal change level (5%) measured by EPI functional MRI studies is much larger than the relative SD (0.15%).
Influence of MRI on PET Performance
The influence of the static 7-T magnetic field and the timevarying radiofrequency waves on the PET system was negligibly
small. Neither shape distortion nor baseline
drift of PET scintillation pulses due to gradient
change and radiofrequency excitation was
observed by analyzing scintillation pulses
captured with a 125-Msps analog-to-digital
converter (Supplemental Fig. 4). No remarkable changes in flood images and energy
spectra were observed after radiofrequency
sequences were applied (Supplemental Figs. 5
and 6). There was only a slight increase in
scattering events caused by radiofrequency
coil insertion. When we compared PET
images of the uniform phantom acquired
with standalone PET and simultaneous
PET/MRI, only an additional attenuation
in PET images was caused by the radiofrequency coil (Fig. 3A). With the radiofrequency coil, 9% less prompt counts were
collected. Interestingly, the spatial resolution of PET shown in the hot-rod phantom
image was better inside the MR scanner
FIGURE 2. (A) Mean signal intensity (left), SNR (center), and uniformity (right) of MR images of
(Fig. 3A) as a result of shortened positron
uniform corn oil phantom. (B) B0 map of a coronal slice for phantom measured before and after
range in the strong magnetic field (29,30).
inserting PET. (C) Echo planar images of phantom and signal intensity before and after inserting
Single-event counting rates of PET were
PET. a.u. 5 arbitrary units.
1312
THE JOURNAL
OF
NUCLEAR MEDICINE • Vol. 57 • No. 8 • August 2016
Downloaded from jnm.snmjournals.org by Dong Soo Lee on November 8, 2016. For personal use only.
perfectly coordinated multiparametric information. The specific uptake of 11C-MET in
U87MG was clearly visualized in PET while
T2-weighted and FLAIR MR images showed
prominent changes in the tumor (Figs. 5A
and 5B; Supplemental Fig. 7). The higher
ADC values in U87MG than the surrounding region added to the confidence of
tumor detection (Fig. 5C).
LN Imaging
A left popliteal LN (node draining from
the injection site) in the 2-h postinjection
MR image resulted in remarkable signal
FIGURE 3. (A) Uniform and hot-rod phantom images and profiles across indicated positions
decrease (negative contrast) compared with
acquired under 2 different conditions: after installing PET outside MRI (solid line) and inserting
that in the MR image acquired before inPET into MRI and running 2D FSE sequence (dashed line). Attenuation due to radiofrequency coil
jection and was in good agreement with
was not corrected. (B) PET single-event counting rate measured under 4 different conditions
64Cu signals from PET (Figs. 6A and 6C),
inside MRI: without running any RF pulse and while running 2D FSE, 3D SPGR, and gradient-echo
EPI sequences. a.u. 5 arbitrary units; RF 5 radiofrequency.
providing solid indication of the LN-specific
uptake of the tracer. The 10-min postinjecmeasured using the intrinsic radioactivity of the LYSO crystal tion PET also showed clear regional activity in the LN, but the 10-min
(blank scan), and there was no significant change found when var- uptake period was not sufficient to generate MR contrast (Fig. 6B);
ious radiofrequency pulse sequences were applied, except for the this is mainly because of the difference in the sensitivities of 2 mogradient-intensive EPI sequence, which yielded a 2% increment in dalities. The ex vivo examination demonstrated the existence of ferric
counting rate (Fig. 3B). The counting rate change would be less iron (Fe31) in the left popliteal LN but not in the right popliteal LN
perceptible in animal studies in which the actual PET counting rates (Supplemental Fig. 8).
are much higher than that in a blank scan.
DISCUSSION
Glucose Metabolism Imaging
Thanks to the sufficiently long axial FOV of our PET system, the
whole mouse brain including the cerebellum and olfactory bulbs
could be assessed in a single bed position with uniform spatial
resolution (Fig. 4). Moreover, high-resolution 7-T MRI offers
excellent anatomic information on the tiny brain structures of the
mouse, enabling us to readily examine the 18F-FDG distribution
in the cerebral cortex and neighboring structures.
Tumor Imaging
Another primary strength of simultaneous PET/MRI is the diversity of PET radiopharmaceuticals and MRI sequences, offering
FIGURE 4. Sagittal (left), transverse (center), and coronal (right) images of 18F-FDG PET/MRI. Positions of coronal sections are denoted
in sagittal and transverse section by i–iv. High uptake in specific brain
regions can be seen including striatum, thalamus, and cerebral cortex.
a.u. 5 arbitrary units; max 5 maximum; min 5 minimum.
One of the most promising applications of SiPM is simultaneous
PET/MR imaging of animal models of human diseases, but there
are several technical challenges. This work was motivated by
the goals of building a small-animal PET/MRI system with fine
spatial resolution, high sensitivity, high temporal stability, and
minimal mutual interference and demonstrating the feasibility of
this system for various animal experiments.
The developed SiPM PET system achieves submillimeter spatial
resolution in the center of the FOV (0.7 mm in-plane using iterative
image reconstruction of a point source in air) desired for rodent brain
imaging and high peak system sensitivity (.3%) desired for dynamic
PET studies with short-lived radiotracers and short time frames (22).
Our new PET system neither significantly degrades the MRI information nor is affected by MRI acquisition. In particular, the minimal
FIGURE 5. (A) Brain tumor images of 11C-MET PET, T2-weighted FSE
MRI, and fused PET/MRI. (B) FLAIR image and fused PET/MRI. (C)
Calculated ADC map along x-direction. a.u. 5 arbitrary units; max 5
maximum; min 5 minimum.
SIMULTANEOUS MULTIPARAMETRIC PET/MRI • Ko et al.
1313
Downloaded from jnm.snmjournals.org by Dong Soo Lee on November 8, 2016. For personal use only.
sensitivity will also be possible via the incorporation of depth-ofinteraction measurement technology in PET detector modules (39).
In addition, incorporating a total variation minimization algorithm
into imaging reconstruction will be useful to improve image quality
by reducing gap artifact (40).
CONCLUSION
FIGURE 6. (A) T2-weighted MR image of lower abdominal area showing popliteal LN before injection. (B) T2-weighted MR and fused PET/
MR images obtained 10 min after injection. Arrow indicates draining
popliteal LN. (C) T2-weighted MR and fused PET/MR images obtained
120 min after injection. Signal attenuation is more obvious than 10-min
postinjection MR image shown in B. a.u. 5 arbitrary units; max 5
maximum; min 5 minimum.
effects of SiPM PET on EPI MR images will allow us to perform
simultaneous PET/functional MRI studies, offering a unique methodology for understanding the relationship between hemodynamic
and neurochemical processes in the brain (31,32). In addition, the
B0 uniformity will enable multiparametric molecular imaging studies
with PET tracers and new emerging MR molecular imaging technologies, such as chemical exchange saturation transfer imaging (33).
Our work shows that simultaneous PET/MR imaging with a
high-performance SiPM PET and an ultra-high-field small-animal
dedicated MR scanner offers synergistic advantages for investigating human disease models in the mouse without significant
compromise in the imaging performances of both modalities. The
highly specific uptake of 11C-MET in the tumor was well correlated
with morphologic changes revealed by high-resolution MR images.
In addition, parametric MR images of cellular hemodynamics
(ADC) incorporated with PET images provide complementary
information needed to understand mechanisms underlying the
structural alterations in the tumor.
Another important application of simultaneous PET/MRI is the
assessment of multimodal imaging probes. Although many interesting
multimodal imaging probes have been suggested (34–36), their spatiotemporal characteristics have been investigated using each imaging
device separately, making the exact correlation and comparison studies impossible. In contrast, our study with 64Cu-NOTA-IO-MAN
showed that high-sensitivity PET images for LN mapping were useful
for confirming the accumulation of iron oxide nanoparticles in the
same target region. The negative T2 or T2* contrast enhanced by
these MRI agents using just MRI may be hard to distinguish from the
inhomogeneity artifacts.
Although the present SiPM-based PET/MRI system showed
excellent feasibility in mouse PET/MRI studies, there is room for
performance improvements. The traditional multiplexing method
used in this study for reducing the number of readout channels (21)
somewhat degrades timing and counting rate performances, which
can be improved using digital SiPM technology or novel sampling
theory (37,38). The improved spatial resolution without the loss of
1314
THE JOURNAL
OF
We developed a multiparametric imager with high physical
performance and good system stability to temperature variation
and demonstrated its feasibility for small-animal experiments,
suggesting its usefulness for investigating in vivo molecular
interactions of metabolites, and cross-validation studies of both
PET and MRI. Although we focused on preclinical studies performed
inside a small-bore 7-T magnet, the technologies developed in
this study can be translated to human PET/MRI scanners with 7-T
or higher-field magnets through further technologic developments and appropriate modifications, which will lead the next
decade of MRI.
DISCLOSURE
The costs of publication of this article were defrayed in part by
the payment of page charges. Therefore, and solely to indicate this
fact, this article is hereby marked “advertisement” in accordance
with 18 USC section 1734. This work was supported in part by
grants from the National Research Foundation of Korea (NRF)
funded by the Korean Ministry of Science, ICT and Future Planning (2008-2003852 and 2014M3C7034000). The funding source
had no involvement in the study design, collection, analysis, or
interpretation. No other potential conflict of interest relevant to
this article was reported.
ACKNOWLEDGMENTS
We thank Seung Won Kim and ChangKi Min for assistance
with animal preparation and MRI data acquisition, and Yun Sang
Park for the design of the 3D printed frame.
REFERENCES
1. Pomper MG, Lee JS. Small animal imaging in drug development. Curr Pharm
Des. 2005;11:3247–3272.
2. Rudin M, Weissleder R. Molecular imaging in drug discovery and development.
Nat Rev Drug Discov. 2003;2:123–131.
3. Kang JH, Lee DS, Peang JC, et al. Development of a sodium/iodide symporter
(NIS)-transgenic mouse for imaging of cardiomyocyte-specific reporter gene
expression. J Nucl Med. 2005;46:479–483.
4. Lee S, Park H. Parametric response mapping of longitudinal PET scans and their
use in detecting changes in Alzheimer’s diseases. Biomed Eng Lett. 2014;4:73–79.
5. Kim MH, Lee YJ, Kang JH. Stem cell monitoring with a direct or indirect labeling
method. Nucl Med Mol Imaging. October 22, 2015 [Epub ahead of print].
6. Pichler BJ, Kolb A, Nagele T, Schlemmer H-P. PET/MRI: paving the way for the
next generation of clinical multimodality imaging applications. J Nucl Med.
2010;51:333–336.
7. Drzezga A, Barthel H, Minoshima S, Sabri O. Potential clinical applications of
PET/MR imaging in neurodegenerative diseases. J Nucl Med. 2014;55:47S–55S.
8. Hildebrandt IJ, Su H, Weber WA. Anesthesia and other considerations for in vivo
imaging of small animals. ILAR J. 2008;49:17–26.
9. Shao Y, Cherry SR, Farahani K, et al. Simultaneous PET and MR imaging. Phys
Med Biol. 1997;42:1965–1970.
10. Raylman RR, Majewski S, Lemieux SK, et al. Simultaneous MRI and PET
imaging of a rat brain. Phys Med Biol. 2006;51:6371–6379.
11. Judenhofer MS, Wehrl HF, Newport DF, et al. Simultaneous PET-MRI: a new
approach for functional and morphological imaging. Nat Med. 2008;14:459–465.
NUCLEAR MEDICINE • Vol. 57 • No. 8 • August 2016
Downloaded from jnm.snmjournals.org by Dong Soo Lee on November 8, 2016. For personal use only.
12. Catana C, Procissi D, Wu Y, et al. Simultaneous in vivo positron emission
tomography and magnetic resonance imaging. Proc Natl Acad Sci USA.
2008;105:3705–3710.
13. Schlemmer H-PW, Pichler BJ, Schmand M, et al. Simultaneous MR/PET imaging of the human brain: feasibility study. Radiology. 2008;248:1028–1035.
14. Delso G, Fürst S, Jakoby B, et al. Performance measurements of the Siemens mMR
integrated whole-body PET/MR scanner. J Nucl Med. 2011;52:1914–1922.
15. Hong SJ, Song IC, Ito M, et al. An investigation into the use of Geiger-mode
solid-state photomultipliers for simultaneous PET and MRI acquisition. IEEE
Trans Nucl Sci. 2008;55:882–888.
16. Schaart DR, Van Dam HT, Seifert S, et al. A novel, SiPM-array-based, monolithic scintillator detector for PET. Phys Med Biol. 2009;54:3501–3512.
17. Kwon SI, Lee JS, Yoon HS, et al. Development of small-animal PET prototype
using silicon photomultiplier (SiPM): initial results of phantom and animal imaging studies. J Nucl Med. 2011;52:572–579.
18. Roncali E, Cherry SR. Application of silicon photomultipliers to positron emission tomography. Ann Biomed Eng. 2011;39:1358–1377.
19. Yoon HS, Ko GB, Kwon SI, et al. Initial results of simultaneous PET/MRI
experiments with an MRI-compatible silicon photomultiplier PET scanner.
J Nucl Med. 2012;53:608–614.
20. Yamamoto S, Watabe T, Watabe H, et al. Simultaneous imaging using Si-PMbased PET and MRI for development of an integrated PET/MRI system. Phys
Med Biol. 2012;57:N1–N13.
21. Ko GB, Yoon HS, Kwon SI, et al. Development of a front-end analog circuit for
multi-channel SiPM readout and performance verification for various PET detector designs. Nucl Instrum Methods Phys Res A. 2013;703:38–44.
22. Ko GB, Kim KY, Yoon HS, et al. Evaluation of a silicon photomultiplier PET
insert for simultaneous PET and MR imaging. Med Phys. 2016;43:72–83.
23. Wehrl HF, Judenhofer MS, Thielscher A, Martirosian P, Schick F, Pichler BJ.
Assessment of MR compatibility of a PET insert developed for simultaneous
multiparametric PET/MR imaging on an animal system operating at 7 T. Magn
Reson Med. 2011;65:269–279.
24. Hong SJ, Kang HG, Ko GB, Song IC, Rhee J-T, Lee JS. SiPM-PET with a short
optical fiber bundle for simultaneous PET-MR imaging. Phys Med Biol.
2012;57:3869–3883.
25. Loening AM, Gambhir SS. AMIDE: a free software tool for multimodality
medical image analysis. Mol Imaging. 2003;2:131–137.
26. Lee JS, Park KS, Lee DS, Lee CW, Chung J-K, Lee MC. Development and
applications of a software for functional image registration (FIRE). Comput
Methods Programs Biomed. 2005;78:157–164.
27. Yang BY, Moon S-H, Seelam SR, et al. Development of a multimodal imaging
probe by encapsulating iron oxide nanoparticles with functionalized amphiphiles
for lymph node imaging. Nanomedicine (Lond). 2015;10:1899–1910.
28. Thorek DLJ, Ulmert D, Diop N-FM, et al. Non-invasive mapping of deep-tissue
lymph nodes in live animals using a multimodal PET/MRI nanoparticle. Nat
Commun. 2014;5:3097.
29. Iida H, Kanno I, Murakami M, et al. A simulation study of a method to reduce
positron annihilation spread distributions using a strong magnetic field in positron emission tomography. IEEE Trans Nucl Sci. 1986;33:597–600.
30. Shah NJ, Herzog H, Weirich C, et al. Effects of magnetic fields of up to 9.4 T on
resolution and contrast of PET images as measured with an MR-brainPET. PLoS
One. 2014;9:e95250.
31. Wehrl HF, Hossain M, Lankes K, et al. Simultaneous PET-MRI reveals brain
function in activated and resting state on metabolic, hemodynamic and multiple
temporal scales. Nat Med. 2013;19:1184–1189.
32. Sander CY, Hooker JM, Catana C, et al. Neurovascular coupling to D2/D3
dopamine receptor occupancy using simultaneous PET/functional MRI. Proc
Natl Acad Sci USA. 2013;110:11169–11174.
33. Cai K, Haris M, Singh A, et al. Magnetic resonance imaging of glutamate. Nat
Med. 2012;18:302–306.
34. Lee H-Y, Li Z, Chen K, et al. PET/MRI dual-modality tumor imaging using
arginine-glycine-aspartic (RGD)-conjugated radiolabeled iron oxide nanoparticles. J Nucl Med. 2008;49:1371–1379.
35. Glaus C, Rossin R, Welch MJ, Bao G. In vivo evaluation of 64Cu-labeled magnetic nanoparticles as a dual-modality PET/MR imaging agent. Bioconjug Chem.
2010;21:715–722.
36. Yang X, Hong H, Grailer JJ, et al. cRGD-functionalized, DOX-conjugated,
and 64Cu-labeled superparamagnetic iron oxide nanoparticles for targeted anticancer drug delivery and PET/MR imaging. Biomaterials. 2011;32:4151–
4160.
37. Haemisch Y, Frach T, Degenhardt C, Thon A. Fully digital arrays of silicon
photomultipliers (dSiPM): a scalable alternative to vacuum photomultiplier tubes
(PMT). Phys Procedia. 2012;37:1546–1560.
38. van den Berg E, Candès E, Chinn G, et al. Single-photon sampling architecture
for solid-state imaging sensors. Proc Natl Acad Sci USA. 2013;110:E2752–
E2761.
39. Ito M, Hong SJ, Lee JS. Positron emission tomography (PET) detectors with
depth-of-interaction (DOI) capability. Biomed Eng Lett. 2011;1:70–81.
40. Son J, Kim SM, Lee JS. A strategy to reduce blocky pattern and contrast loss in
emission tomography reconstruction with reduced angular sampling and total
variation minimization. Biomed Eng Lett. 2014;4:362–369.
SIMULTANEOUS MULTIPARAMETRIC PET/MRI • Ko et al.
1315
Downloaded from jnm.snmjournals.org by Dong Soo Lee on November 8, 2016. For personal use only.
Simultaneous Multiparametric PET/MRI with Silicon Photomultiplier PET and
Ultra-High-Field MRI for Small-Animal Imaging
Guen Bae Ko, Hyun Suk Yoon, Kyeong Yun Kim, Min Sun Lee, Bo Yeun Yang, Jae Min Jeong, Dong Soo Lee, In
Chan Song, Seok-ki Kim, Daehong Kim and Jae Sung Lee
J Nucl Med. 2016;57:1309-1315.
Published online: April 14, 2016.
Doi: 10.2967/jnumed.115.170019
This article and updated information are available at:
http://jnm.snmjournals.org/content/57/8/1309
Information about reproducing figures, tables, or other portions of this article can be found online at:
http://jnm.snmjournals.org/site/misc/permission.xhtml
Information about subscriptions to JNM can be found at:
http://jnm.snmjournals.org/site/subscriptions/online.xhtml
The Journal of Nuclear Medicine is published monthly.
SNMMI | Society of Nuclear Medicine and Molecular Imaging
1850 Samuel Morse Drive, Reston, VA 20190.
(Print ISSN: 0161-5505, Online ISSN: 2159-662X)
© Copyright 2016 SNMMI; all rights reserved.